Computer Vision Demo / Workshop

This demo / workshop is designed to showcase OpenShift as a platform for developing and operationalizing AI/ML applications.

Specifically it focuses on a computer vision implementation and covers a workflow for custom training and serving for integration with other applications and services. It uses many tools in the Red Hat ecosystem as well as 3rd party applications and services.

Here is a video series explaining the workshop and a runthrough of the setup and use cases.

Prerequisites

CLI tools

• bash, git
• oc - Download mac, linux, windows

Infrastructure

• OpenShift 4.12+
  • role: cluster-admin
• Nvidia GPU hardware
• Internet access

Optional

• OpenShift Dev Spaces 3.8.0+
• AWS (auto scaling)

Red Hat Demo Platform Options (Tested)

• AWS with OpenShift Open Environment
  • 1 x Control Plane - m5.4xlarge
  • 1 x Workers - m5.2xlarge
• One Node OpenShift
  • 1 x Control Plane - m5.4xlarge
• MLOps Demo: Data Science & Edge Practice

Quickstart

You can use OpenShift Dev Spaces for AI / ML or your local machine to explore this demo.

# cluster tasks
scripts/bootstrap.sh

# namespace only tasks
oc new-project let-me-fly-away
oc apply -k gitops/user-components

# create a new pipeline run
oc create -f components/pipelines/manifests/pipeline-run.yaml

Workshop Instructions

See docs/WORKSHOP.md

Additional Resources

• OpenShift Dev Spaces for AI / ML